SlideShare a Scribd company logo
Wednesday, 27th March
A Post-Hoc Categorisation
of Predictive Models
John Mitros
University College Dublin
ioannis.mitros@insight-centre.org
Outline
• Introduc)on
• Overview of interpretability/explainability
• Post-hoc approaches for interpretability
• Common themes, connec)ng ideas, general picture
• Not an exhaus)ve survey of all the literature body
• Open challenges and possible future direc)ons
• Examples and use cases
• Recent approaches
2
Explainable AI
3
Image from https://xaitutorial2019.github.io/
Explainable vs. Interpretable
• Explainable ML:
• Post-hoc analysis of black box models
• Interpretable ML:
• Intrinsically interpretable a.k.a transparent
4Rudin, C. & Ertekin, Ş. Math. Prog. Comp. (2018) 10: 659. hNps://doi.org/10.1007/s12532-018-0143-8
Interpretability
• It is inherently a mul/faceted no/on whose meaning changes according to
the different applicability scenarios
• Interpretability needs to answer what the model has learned and why it
came to that conclusion
• Defini/on of interpretability:
• “interpretability is the degree to which a human can understand the cause of a
decision” (Miller 2017)
5
Interpretability
• Defini&on of interpretability:
• “interpretability is the degree to which a machine can explain the cause of a decision
into coherent logical arguments”
• inherently it involves a bijec&ve process from input to output and vice versa, where
the intermediate steps are transparent to the end user
!" " # → % &ℎ() " % → #
• logical fallacies should be avoided
6
Scope of Interpretability
7
Lipton, Z. C. 2016. The Mythos of Model Interpretability. ICML Workshop on Human Interpretability in Machine Learning
(WHI 2016), New York, NY
Scope of Interpretability
8
Hierarchical Structure of Interpretability
9
Examples of Post-hoc Explana2ons
10
Chen, C.; Li, O.; Barne1, A.; Su, J.; and Rudin, C. 2018. This looks like that: Deep learning for interpretable image
recogniHon. ICML
Group A
What has the model
learned?
(holisHc or modular level)
Model
Specific
Examples of Post-hoc Explana2ons
11
Rudin, C., and Ertekin, S ̧. 2018. Learning customized and op>mized lists of rules with mathema>cal programming.
Mathema'cal Programming Computa'on 10(4):659–702
Group A
What has the model
learned?
(holis>c or modular level)
Model
Specific
Examples of Post-hoc Explana2ons
12
Montavon, G.; Lapuschkin, S.; Binder, A.; Samek, W.; and Mu ̈ller, K.-R. 2017. Explaining nonlinear classificaIon de- cisions
with deep Taylor decomposiIon. Pa#ern Recogni- .on 65:211–222.
Group A
What has the model
learned?
(holisIc or modular level)
Model
Specific
Examples of Post-hoc Explana2ons
13
Ribeiro, M. T.; Singh, S.; and Guestrin, C. 2016. ”Why Should I Trust You?”: Explaining the PredicJons of Any Classifier.
ACM KDD
Group A
What has the model
learned?
(holisJc or modular level)
Model
AgnosJc
Examples of Post-hoc Explana2ons
14Ribeiro, M. T.; Singh, S.; and Guestrin, C. 2018. Anchors: High-Precision Model-Agnostic Explanations. AAAI Press 32:1527–
Group A
What has the model
learned?
(holisOc or modular level)
Model
AgnosOc
Examples of Post-hoc Explana2ons
15
Henelius, A.; Puolama ̈ki, K.; and Ukkonen, A. 2017. Interpre?ng Classifiers through ADribute Interac?ons in Datasets.
ICML
Group A
What has the model
learned?
(holis?c or modular level)
Model
Agnos?c
General Concepts & Methods
• Rule Sets
• Sensi+vity Analysis
• Induc+ve Logic/Programming
• Recently:
• Counterfactuals
• Adversarial approaches
• Game theory
16
Interes'ng Direc'ons
17
Interes'ng Direc'ons
18
Open Challenges
• No formal agreed upon defini1on
• The no1on of interpretability seems to be an ill-defined term?
• Having agreed upon defini1on avoids reinven1ng the wheel
• Easier to built upon and contribute to prior work
• Rigorous, agreed upon evalua1on metrics
• Clear dis1nc1on of human vs. machine based evalua1on metrics
• Provide a clear picture of what is working and what needs improvement
19
Open Challenges
• Stochas(c nature of the models, different random seeds lead to different
outcomes for the same models
• P Henderson, R Islam, P Bachman, J Pineau, Deep Reinforcement Learning That
MaAers, AAAI 20018
• Models are built on assump(ons à = f( )
• When do they break and how?
20
Open Challenges
• Humans are great storytellers/story makers
• Memory championship à Method of loci
• Often humans create stories from small indications which rely upon in order to
build explanations
• These explanations might not have any relation with the underlying actual model
• How to avoid specific cognitive biases?
• Framing effect
• Focusing effect
• Illusory correlation 21
Open Challenges
22
Open Challenges
23
Open Challenges
• Saliency maps can be misleading (Olah et al., 2018)
• Models are uncalibrated
• Need for more transparent approaches
• Bringing another to interpret the exisCng 24
References
1. Bodenhofer, Ulrich and Bauer, Peter. Towards an Axiomatic Treatment of
Interpretability, Fuzzy Systems, 2000.
2. Olah, Chris and Satyanarayan, Arvind. The Building Blocks of Interpretability, Distill.pub,
2018.
3. Zadrozny, Bianca and Elkan, Charles. Obtaining calibrated probability estimates from
decision trees and naive bayesian classifiers. In ICML, pp. 609–616, 2001.
4. Zadrozny, Bianca and Elkan, Charles. Transforming classifier scores into accurate
multiclass probability estimates. In KDD, pp. 694–699, 2002.
5. Naeini, Mahdi Pakdaman, Cooper, Gregory F, and Hauskrecht, Milos. Obtaining well
calibrated probabilities using bayesian binning. In AAAI, pp. 2901, 2015.
6. Platt, John et al. Probabilistic outputs for support vector machines and comparisons to
regularized likelihood methods. Advances in large margin classifiers, 10(3): 61–74, 1999.
7. Guo, Chuan and Pleiss, Geoff and Sun, Yu and Weinberger, Kilian Q. On Calibration of
Modern Neural Networks. In ICML 2017.
25
THANKS!
Ques-ons?
26
Preprint: https://arxiv.org/abs/1904.02495
ioannis.mitros@insight-centre.org

More Related Content

What's hot

Semantics-based Graph Approach to Complex Question-Answering
Semantics-based Graph Approach to Complex Question-AnsweringSemantics-based Graph Approach to Complex Question-Answering
Semantics-based Graph Approach to Complex Question-Answering
Jinho Choi
 
A Review of Distributional models of word meaning (Lenci, 2018)
A Review of Distributional models of word meaning (Lenci, 2018)A Review of Distributional models of word meaning (Lenci, 2018)
A Review of Distributional models of word meaning (Lenci, 2018)
Shuntaro Yada
 
Introduction to gre quant
Introduction to gre quantIntroduction to gre quant
Introduction to gre quant
Crack The Quant
 
Marking breakdown
Marking breakdownMarking breakdown
Marking breakdown
Queensbury 6th Form
 
Cara apakah a raven matriks uji
Cara apakah a raven matriks ujiCara apakah a raven matriks uji
Cara apakah a raven matriks uji
Agel Hatmaja
 
Relationship of knowledge to learn in programming methodology and evaluation ...
Relationship of knowledge to learn in programming methodology and evaluation ...Relationship of knowledge to learn in programming methodology and evaluation ...
Relationship of knowledge to learn in programming methodology and evaluation ...
Technological Ecosystems for Enhancing Multiculturality
 
Computational methods couurseout line
Computational methods couurseout lineComputational methods couurseout line
Computational methods couurseout line
Temesgen Geta
 
OMSU vs. EU comparative curricula study
OMSU vs. EU comparative curricula studyOMSU vs. EU comparative curricula study
OMSU vs. EU comparative curricula study
metamath
 
Flowchart and concept maps
Flowchart and concept mapsFlowchart and concept maps
Flowchart and concept maps
Shruti Nimbkar
 
Geometry thinking
Geometry thinkingGeometry thinking
M phil
M philM phil
M phil
peeroz
 
GRE Presentation
GRE PresentationGRE Presentation
GRE Presentation
Western Academy
 
Montpellier
MontpellierMontpellier
Montpellier
Ben Bolker
 
Applied logic: A mastery learning approach
Applied logic: A mastery learning approachApplied logic: A mastery learning approach
Applied logic: A mastery learning approach
john6938
 
Numerical Analysis And Linear Algebra
Numerical Analysis And Linear AlgebraNumerical Analysis And Linear Algebra
Numerical Analysis And Linear Algebra
Ghulam Murtaza
 
Pattern for gate 2014
Pattern for gate 2014Pattern for gate 2014
Pattern for gate 2014
sakshij91
 
先進網際服務系統1008
先進網際服務系統1008先進網際服務系統1008
先進網際服務系統1008
youji0811
 
Lec 6 learning
Lec 6 learningLec 6 learning
Lec 6 learning
Eyob Sisay
 
Sample prac exam2013
Sample prac exam2013Sample prac exam2013
Sample prac exam2013
hccit
 
computational science & engineering seminar, 16 oct 2013
computational science & engineering seminar, 16 oct 2013computational science & engineering seminar, 16 oct 2013
computational science & engineering seminar, 16 oct 2013
Ben Bolker
 

What's hot (20)

Semantics-based Graph Approach to Complex Question-Answering
Semantics-based Graph Approach to Complex Question-AnsweringSemantics-based Graph Approach to Complex Question-Answering
Semantics-based Graph Approach to Complex Question-Answering
 
A Review of Distributional models of word meaning (Lenci, 2018)
A Review of Distributional models of word meaning (Lenci, 2018)A Review of Distributional models of word meaning (Lenci, 2018)
A Review of Distributional models of word meaning (Lenci, 2018)
 
Introduction to gre quant
Introduction to gre quantIntroduction to gre quant
Introduction to gre quant
 
Marking breakdown
Marking breakdownMarking breakdown
Marking breakdown
 
Cara apakah a raven matriks uji
Cara apakah a raven matriks ujiCara apakah a raven matriks uji
Cara apakah a raven matriks uji
 
Relationship of knowledge to learn in programming methodology and evaluation ...
Relationship of knowledge to learn in programming methodology and evaluation ...Relationship of knowledge to learn in programming methodology and evaluation ...
Relationship of knowledge to learn in programming methodology and evaluation ...
 
Computational methods couurseout line
Computational methods couurseout lineComputational methods couurseout line
Computational methods couurseout line
 
OMSU vs. EU comparative curricula study
OMSU vs. EU comparative curricula studyOMSU vs. EU comparative curricula study
OMSU vs. EU comparative curricula study
 
Flowchart and concept maps
Flowchart and concept mapsFlowchart and concept maps
Flowchart and concept maps
 
Geometry thinking
Geometry thinkingGeometry thinking
Geometry thinking
 
M phil
M philM phil
M phil
 
GRE Presentation
GRE PresentationGRE Presentation
GRE Presentation
 
Montpellier
MontpellierMontpellier
Montpellier
 
Applied logic: A mastery learning approach
Applied logic: A mastery learning approachApplied logic: A mastery learning approach
Applied logic: A mastery learning approach
 
Numerical Analysis And Linear Algebra
Numerical Analysis And Linear AlgebraNumerical Analysis And Linear Algebra
Numerical Analysis And Linear Algebra
 
Pattern for gate 2014
Pattern for gate 2014Pattern for gate 2014
Pattern for gate 2014
 
先進網際服務系統1008
先進網際服務系統1008先進網際服務系統1008
先進網際服務系統1008
 
Lec 6 learning
Lec 6 learningLec 6 learning
Lec 6 learning
 
Sample prac exam2013
Sample prac exam2013Sample prac exam2013
Sample prac exam2013
 
computational science & engineering seminar, 16 oct 2013
computational science & engineering seminar, 16 oct 2013computational science & engineering seminar, 16 oct 2013
computational science & engineering seminar, 16 oct 2013
 

Similar to A Categorisation of Post-hoc Explanations for Predictive Models

0 CPE105 ORIENTATION.ppsx
0 CPE105 ORIENTATION.ppsx0 CPE105 ORIENTATION.ppsx
0 CPE105 ORIENTATION.ppsx
SioMairice11
 
Discussant EARLI sig 27
Discussant EARLI sig 27Discussant EARLI sig 27
Discussant EARLI sig 27
Christian Bokhove
 
​​Explainability in AI and Recommender systems: let’s make it interactive!
​​Explainability in AI and Recommender systems: let’s make it interactive!​​Explainability in AI and Recommender systems: let’s make it interactive!
​​Explainability in AI and Recommender systems: let’s make it interactive!
Eindhoven University of Technology / JADS
 
Inverse Modeling for Cognitive Science "in the Wild"
Inverse Modeling for Cognitive Science "in the Wild"Inverse Modeling for Cognitive Science "in the Wild"
Inverse Modeling for Cognitive Science "in the Wild"
Aalto University
 
Semantic Techniques for Enabling Knowledge Reuse in Conceptual Modelling
Semantic Techniques for Enabling Knowledge Reuse in Conceptual ModellingSemantic Techniques for Enabling Knowledge Reuse in Conceptual Modelling
Semantic Techniques for Enabling Knowledge Reuse in Conceptual Modelling
Oscar Corcho
 
HCI Research as Problem-Solving
HCI Research as Problem-SolvingHCI Research as Problem-Solving
HCI Research as Problem-Solving
Minjoon Kim
 
Course contents
Course contentsCourse contents
Course contents
Syed Shah
 
Modelling presentation
Modelling presentationModelling presentation
Modelling presentation
John Liakos
 
ML Interpretability Inside Out
ML Interpretability Inside OutML Interpretability Inside Out
ML Interpretability Inside Out
Mara Graziani
 
2015EDM: A Framework for Multifaceted Evaluation of Student Models (Polygon)
2015EDM: A Framework for Multifaceted Evaluation of Student Models (Polygon)2015EDM: A Framework for Multifaceted Evaluation of Student Models (Polygon)
2015EDM: A Framework for Multifaceted Evaluation of Student Models (Polygon)
Yun Huang
 
1_Q2-PRACTICAL-RESEARCH.pptx
1_Q2-PRACTICAL-RESEARCH.pptx1_Q2-PRACTICAL-RESEARCH.pptx
1_Q2-PRACTICAL-RESEARCH.pptx
GeraldRefil3
 
Seminar University of Loughborough: Using technology to support mathematics e...
Seminar University of Loughborough: Using technology to support mathematics e...Seminar University of Loughborough: Using technology to support mathematics e...
Seminar University of Loughborough: Using technology to support mathematics e...
Christian Bokhove
 
Whether simulation models that fall under the information systems category ad...
Whether simulation models that fall under the information systems category ad...Whether simulation models that fall under the information systems category ad...
Whether simulation models that fall under the information systems category ad...
Elisavet Andrikopoulou
 
Learning to learn Model Behavior: How to use "human-in-the-loop" to explain d...
Learning to learn Model Behavior: How to use "human-in-the-loop" to explain d...Learning to learn Model Behavior: How to use "human-in-the-loop" to explain d...
Learning to learn Model Behavior: How to use "human-in-the-loop" to explain d...
IDEAS - Int'l Data Engineering and Science Association
 
eli2012-learning-analytics
eli2012-learning-analyticseli2012-learning-analytics
eli2012-learning-analytics
Simon Buckingham Shum
 
Apmp brazil oct 2017
Apmp brazil oct 2017Apmp brazil oct 2017
Apmp brazil oct 2017
Brendan Larvor
 
STRUCTURAL ANALYSIS NINTH EDITION R. C. HIBBELER
STRUCTURAL ANALYSIS NINTH EDITION R. C. HIBBELERSTRUCTURAL ANALYSIS NINTH EDITION R. C. HIBBELER
STRUCTURAL ANALYSIS NINTH EDITION R. C. HIBBELER
Bahzad5
 
2018 syllabus
2018 syllabus2018 syllabus
2018 syllabus
sunilSunilmech345
 
Pedagogic Enquiry Presentation - Threshold Concepts in Statistics as a Discip...
Pedagogic Enquiry Presentation - Threshold Concepts in Statistics as a Discip...Pedagogic Enquiry Presentation - Threshold Concepts in Statistics as a Discip...
Pedagogic Enquiry Presentation - Threshold Concepts in Statistics as a Discip...
Richard Diamond
 
Human in the loop: Bayesian Rules Enabling Explainable AI
Human in the loop: Bayesian Rules Enabling Explainable AIHuman in the loop: Bayesian Rules Enabling Explainable AI
Human in the loop: Bayesian Rules Enabling Explainable AI
Pramit Choudhary
 

Similar to A Categorisation of Post-hoc Explanations for Predictive Models (20)

0 CPE105 ORIENTATION.ppsx
0 CPE105 ORIENTATION.ppsx0 CPE105 ORIENTATION.ppsx
0 CPE105 ORIENTATION.ppsx
 
Discussant EARLI sig 27
Discussant EARLI sig 27Discussant EARLI sig 27
Discussant EARLI sig 27
 
​​Explainability in AI and Recommender systems: let’s make it interactive!
​​Explainability in AI and Recommender systems: let’s make it interactive!​​Explainability in AI and Recommender systems: let’s make it interactive!
​​Explainability in AI and Recommender systems: let’s make it interactive!
 
Inverse Modeling for Cognitive Science "in the Wild"
Inverse Modeling for Cognitive Science "in the Wild"Inverse Modeling for Cognitive Science "in the Wild"
Inverse Modeling for Cognitive Science "in the Wild"
 
Semantic Techniques for Enabling Knowledge Reuse in Conceptual Modelling
Semantic Techniques for Enabling Knowledge Reuse in Conceptual ModellingSemantic Techniques for Enabling Knowledge Reuse in Conceptual Modelling
Semantic Techniques for Enabling Knowledge Reuse in Conceptual Modelling
 
HCI Research as Problem-Solving
HCI Research as Problem-SolvingHCI Research as Problem-Solving
HCI Research as Problem-Solving
 
Course contents
Course contentsCourse contents
Course contents
 
Modelling presentation
Modelling presentationModelling presentation
Modelling presentation
 
ML Interpretability Inside Out
ML Interpretability Inside OutML Interpretability Inside Out
ML Interpretability Inside Out
 
2015EDM: A Framework for Multifaceted Evaluation of Student Models (Polygon)
2015EDM: A Framework for Multifaceted Evaluation of Student Models (Polygon)2015EDM: A Framework for Multifaceted Evaluation of Student Models (Polygon)
2015EDM: A Framework for Multifaceted Evaluation of Student Models (Polygon)
 
1_Q2-PRACTICAL-RESEARCH.pptx
1_Q2-PRACTICAL-RESEARCH.pptx1_Q2-PRACTICAL-RESEARCH.pptx
1_Q2-PRACTICAL-RESEARCH.pptx
 
Seminar University of Loughborough: Using technology to support mathematics e...
Seminar University of Loughborough: Using technology to support mathematics e...Seminar University of Loughborough: Using technology to support mathematics e...
Seminar University of Loughborough: Using technology to support mathematics e...
 
Whether simulation models that fall under the information systems category ad...
Whether simulation models that fall under the information systems category ad...Whether simulation models that fall under the information systems category ad...
Whether simulation models that fall under the information systems category ad...
 
Learning to learn Model Behavior: How to use "human-in-the-loop" to explain d...
Learning to learn Model Behavior: How to use "human-in-the-loop" to explain d...Learning to learn Model Behavior: How to use "human-in-the-loop" to explain d...
Learning to learn Model Behavior: How to use "human-in-the-loop" to explain d...
 
eli2012-learning-analytics
eli2012-learning-analyticseli2012-learning-analytics
eli2012-learning-analytics
 
Apmp brazil oct 2017
Apmp brazil oct 2017Apmp brazil oct 2017
Apmp brazil oct 2017
 
STRUCTURAL ANALYSIS NINTH EDITION R. C. HIBBELER
STRUCTURAL ANALYSIS NINTH EDITION R. C. HIBBELERSTRUCTURAL ANALYSIS NINTH EDITION R. C. HIBBELER
STRUCTURAL ANALYSIS NINTH EDITION R. C. HIBBELER
 
2018 syllabus
2018 syllabus2018 syllabus
2018 syllabus
 
Pedagogic Enquiry Presentation - Threshold Concepts in Statistics as a Discip...
Pedagogic Enquiry Presentation - Threshold Concepts in Statistics as a Discip...Pedagogic Enquiry Presentation - Threshold Concepts in Statistics as a Discip...
Pedagogic Enquiry Presentation - Threshold Concepts in Statistics as a Discip...
 
Human in the loop: Bayesian Rules Enabling Explainable AI
Human in the loop: Bayesian Rules Enabling Explainable AIHuman in the loop: Bayesian Rules Enabling Explainable AI
Human in the loop: Bayesian Rules Enabling Explainable AI
 

Recently uploaded

Experts live - Improving user adoption with AI
Experts live - Improving user adoption with AIExperts live - Improving user adoption with AI
Experts live - Improving user adoption with AI
jitskeb
 
The Ipsos - AI - Monitor 2024 Report.pdf
The  Ipsos - AI - Monitor 2024 Report.pdfThe  Ipsos - AI - Monitor 2024 Report.pdf
The Ipsos - AI - Monitor 2024 Report.pdf
Social Samosa
 
Intelligence supported media monitoring in veterinary medicine
Intelligence supported media monitoring in veterinary medicineIntelligence supported media monitoring in veterinary medicine
Intelligence supported media monitoring in veterinary medicine
AndrzejJarynowski
 
一比一原版(Glasgow毕业证书)格拉斯哥大学毕业证如何办理
一比一原版(Glasgow毕业证书)格拉斯哥大学毕业证如何办理一比一原版(Glasgow毕业证书)格拉斯哥大学毕业证如何办理
一比一原版(Glasgow毕业证书)格拉斯哥大学毕业证如何办理
g4dpvqap0
 
Everything you wanted to know about LIHTC
Everything you wanted to know about LIHTCEverything you wanted to know about LIHTC
Everything you wanted to know about LIHTC
Roger Valdez
 
State of Artificial intelligence Report 2023
State of Artificial intelligence Report 2023State of Artificial intelligence Report 2023
State of Artificial intelligence Report 2023
kuntobimo2016
 
4th Modern Marketing Reckoner by MMA Global India & Group M: 60+ experts on W...
4th Modern Marketing Reckoner by MMA Global India & Group M: 60+ experts on W...4th Modern Marketing Reckoner by MMA Global India & Group M: 60+ experts on W...
4th Modern Marketing Reckoner by MMA Global India & Group M: 60+ experts on W...
Social Samosa
 
一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理
一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理
一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理
74nqk8xf
 
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
Timothy Spann
 
一比一原版(牛布毕业证书)牛津布鲁克斯大学毕业证如何办理
一比一原版(牛布毕业证书)牛津布鲁克斯大学毕业证如何办理一比一原版(牛布毕业证书)牛津布鲁克斯大学毕业证如何办理
一比一原版(牛布毕业证书)牛津布鲁克斯大学毕业证如何办理
74nqk8xf
 
一比一原版(UO毕业证)渥太华大学毕业证如何办理
一比一原版(UO毕业证)渥太华大学毕业证如何办理一比一原版(UO毕业证)渥太华大学毕业证如何办理
一比一原版(UO毕业证)渥太华大学毕业证如何办理
aqzctr7x
 
Analysis insight about a Flyball dog competition team's performance
Analysis insight about a Flyball dog competition team's performanceAnalysis insight about a Flyball dog competition team's performance
Analysis insight about a Flyball dog competition team's performance
roli9797
 
一比一原版(GWU,GW文凭证书)乔治·华盛顿大学毕业证如何办理
一比一原版(GWU,GW文凭证书)乔治·华盛顿大学毕业证如何办理一比一原版(GWU,GW文凭证书)乔治·华盛顿大学毕业证如何办理
一比一原版(GWU,GW文凭证书)乔治·华盛顿大学毕业证如何办理
bopyb
 
Palo Alto Cortex XDR presentation .......
Palo Alto Cortex XDR presentation .......Palo Alto Cortex XDR presentation .......
Palo Alto Cortex XDR presentation .......
Sachin Paul
 
一比一原版(UMN文凭证书)明尼苏达大学毕业证如何办理
一比一原版(UMN文凭证书)明尼苏达大学毕业证如何办理一比一原版(UMN文凭证书)明尼苏达大学毕业证如何办理
一比一原版(UMN文凭证书)明尼苏达大学毕业证如何办理
nyfuhyz
 
06-12-2024-BudapestDataForum-BuildingReal-timePipelineswithFLaNK AIM
06-12-2024-BudapestDataForum-BuildingReal-timePipelineswithFLaNK AIM06-12-2024-BudapestDataForum-BuildingReal-timePipelineswithFLaNK AIM
06-12-2024-BudapestDataForum-BuildingReal-timePipelineswithFLaNK AIM
Timothy Spann
 
一比一原版(Chester毕业证书)切斯特大学毕业证如何办理
一比一原版(Chester毕业证书)切斯特大学毕业证如何办理一比一原版(Chester毕业证书)切斯特大学毕业证如何办理
一比一原版(Chester毕业证书)切斯特大学毕业证如何办理
74nqk8xf
 
Influence of Marketing Strategy and Market Competition on Business Plan
Influence of Marketing Strategy and Market Competition on Business PlanInfluence of Marketing Strategy and Market Competition on Business Plan
Influence of Marketing Strategy and Market Competition on Business Plan
jerlynmaetalle
 
办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样
办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样
办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样
apvysm8
 
Udemy_2024_Global_Learning_Skills_Trends_Report (1).pdf
Udemy_2024_Global_Learning_Skills_Trends_Report (1).pdfUdemy_2024_Global_Learning_Skills_Trends_Report (1).pdf
Udemy_2024_Global_Learning_Skills_Trends_Report (1).pdf
Fernanda Palhano
 

Recently uploaded (20)

Experts live - Improving user adoption with AI
Experts live - Improving user adoption with AIExperts live - Improving user adoption with AI
Experts live - Improving user adoption with AI
 
The Ipsos - AI - Monitor 2024 Report.pdf
The  Ipsos - AI - Monitor 2024 Report.pdfThe  Ipsos - AI - Monitor 2024 Report.pdf
The Ipsos - AI - Monitor 2024 Report.pdf
 
Intelligence supported media monitoring in veterinary medicine
Intelligence supported media monitoring in veterinary medicineIntelligence supported media monitoring in veterinary medicine
Intelligence supported media monitoring in veterinary medicine
 
一比一原版(Glasgow毕业证书)格拉斯哥大学毕业证如何办理
一比一原版(Glasgow毕业证书)格拉斯哥大学毕业证如何办理一比一原版(Glasgow毕业证书)格拉斯哥大学毕业证如何办理
一比一原版(Glasgow毕业证书)格拉斯哥大学毕业证如何办理
 
Everything you wanted to know about LIHTC
Everything you wanted to know about LIHTCEverything you wanted to know about LIHTC
Everything you wanted to know about LIHTC
 
State of Artificial intelligence Report 2023
State of Artificial intelligence Report 2023State of Artificial intelligence Report 2023
State of Artificial intelligence Report 2023
 
4th Modern Marketing Reckoner by MMA Global India & Group M: 60+ experts on W...
4th Modern Marketing Reckoner by MMA Global India & Group M: 60+ experts on W...4th Modern Marketing Reckoner by MMA Global India & Group M: 60+ experts on W...
4th Modern Marketing Reckoner by MMA Global India & Group M: 60+ experts on W...
 
一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理
一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理
一比一原版(Coventry毕业证书)考文垂大学毕业证如何办理
 
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
 
一比一原版(牛布毕业证书)牛津布鲁克斯大学毕业证如何办理
一比一原版(牛布毕业证书)牛津布鲁克斯大学毕业证如何办理一比一原版(牛布毕业证书)牛津布鲁克斯大学毕业证如何办理
一比一原版(牛布毕业证书)牛津布鲁克斯大学毕业证如何办理
 
一比一原版(UO毕业证)渥太华大学毕业证如何办理
一比一原版(UO毕业证)渥太华大学毕业证如何办理一比一原版(UO毕业证)渥太华大学毕业证如何办理
一比一原版(UO毕业证)渥太华大学毕业证如何办理
 
Analysis insight about a Flyball dog competition team's performance
Analysis insight about a Flyball dog competition team's performanceAnalysis insight about a Flyball dog competition team's performance
Analysis insight about a Flyball dog competition team's performance
 
一比一原版(GWU,GW文凭证书)乔治·华盛顿大学毕业证如何办理
一比一原版(GWU,GW文凭证书)乔治·华盛顿大学毕业证如何办理一比一原版(GWU,GW文凭证书)乔治·华盛顿大学毕业证如何办理
一比一原版(GWU,GW文凭证书)乔治·华盛顿大学毕业证如何办理
 
Palo Alto Cortex XDR presentation .......
Palo Alto Cortex XDR presentation .......Palo Alto Cortex XDR presentation .......
Palo Alto Cortex XDR presentation .......
 
一比一原版(UMN文凭证书)明尼苏达大学毕业证如何办理
一比一原版(UMN文凭证书)明尼苏达大学毕业证如何办理一比一原版(UMN文凭证书)明尼苏达大学毕业证如何办理
一比一原版(UMN文凭证书)明尼苏达大学毕业证如何办理
 
06-12-2024-BudapestDataForum-BuildingReal-timePipelineswithFLaNK AIM
06-12-2024-BudapestDataForum-BuildingReal-timePipelineswithFLaNK AIM06-12-2024-BudapestDataForum-BuildingReal-timePipelineswithFLaNK AIM
06-12-2024-BudapestDataForum-BuildingReal-timePipelineswithFLaNK AIM
 
一比一原版(Chester毕业证书)切斯特大学毕业证如何办理
一比一原版(Chester毕业证书)切斯特大学毕业证如何办理一比一原版(Chester毕业证书)切斯特大学毕业证如何办理
一比一原版(Chester毕业证书)切斯特大学毕业证如何办理
 
Influence of Marketing Strategy and Market Competition on Business Plan
Influence of Marketing Strategy and Market Competition on Business PlanInfluence of Marketing Strategy and Market Competition on Business Plan
Influence of Marketing Strategy and Market Competition on Business Plan
 
办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样
办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样
办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样
 
Udemy_2024_Global_Learning_Skills_Trends_Report (1).pdf
Udemy_2024_Global_Learning_Skills_Trends_Report (1).pdfUdemy_2024_Global_Learning_Skills_Trends_Report (1).pdf
Udemy_2024_Global_Learning_Skills_Trends_Report (1).pdf
 

A Categorisation of Post-hoc Explanations for Predictive Models

  • 1. Wednesday, 27th March A Post-Hoc Categorisation of Predictive Models John Mitros University College Dublin ioannis.mitros@insight-centre.org
  • 2. Outline • Introduc)on • Overview of interpretability/explainability • Post-hoc approaches for interpretability • Common themes, connec)ng ideas, general picture • Not an exhaus)ve survey of all the literature body • Open challenges and possible future direc)ons • Examples and use cases • Recent approaches 2
  • 3. Explainable AI 3 Image from https://xaitutorial2019.github.io/
  • 4. Explainable vs. Interpretable • Explainable ML: • Post-hoc analysis of black box models • Interpretable ML: • Intrinsically interpretable a.k.a transparent 4Rudin, C. & Ertekin, Ş. Math. Prog. Comp. (2018) 10: 659. hNps://doi.org/10.1007/s12532-018-0143-8
  • 5. Interpretability • It is inherently a mul/faceted no/on whose meaning changes according to the different applicability scenarios • Interpretability needs to answer what the model has learned and why it came to that conclusion • Defini/on of interpretability: • “interpretability is the degree to which a human can understand the cause of a decision” (Miller 2017) 5
  • 6. Interpretability • Defini&on of interpretability: • “interpretability is the degree to which a machine can explain the cause of a decision into coherent logical arguments” • inherently it involves a bijec&ve process from input to output and vice versa, where the intermediate steps are transparent to the end user !" " # → % &ℎ() " % → # • logical fallacies should be avoided 6
  • 7. Scope of Interpretability 7 Lipton, Z. C. 2016. The Mythos of Model Interpretability. ICML Workshop on Human Interpretability in Machine Learning (WHI 2016), New York, NY
  • 9. Hierarchical Structure of Interpretability 9
  • 10. Examples of Post-hoc Explana2ons 10 Chen, C.; Li, O.; Barne1, A.; Su, J.; and Rudin, C. 2018. This looks like that: Deep learning for interpretable image recogniHon. ICML Group A What has the model learned? (holisHc or modular level) Model Specific
  • 11. Examples of Post-hoc Explana2ons 11 Rudin, C., and Ertekin, S ̧. 2018. Learning customized and op>mized lists of rules with mathema>cal programming. Mathema'cal Programming Computa'on 10(4):659–702 Group A What has the model learned? (holis>c or modular level) Model Specific
  • 12. Examples of Post-hoc Explana2ons 12 Montavon, G.; Lapuschkin, S.; Binder, A.; Samek, W.; and Mu ̈ller, K.-R. 2017. Explaining nonlinear classificaIon de- cisions with deep Taylor decomposiIon. Pa#ern Recogni- .on 65:211–222. Group A What has the model learned? (holisIc or modular level) Model Specific
  • 13. Examples of Post-hoc Explana2ons 13 Ribeiro, M. T.; Singh, S.; and Guestrin, C. 2016. ”Why Should I Trust You?”: Explaining the PredicJons of Any Classifier. ACM KDD Group A What has the model learned? (holisJc or modular level) Model AgnosJc
  • 14. Examples of Post-hoc Explana2ons 14Ribeiro, M. T.; Singh, S.; and Guestrin, C. 2018. Anchors: High-Precision Model-Agnostic Explanations. AAAI Press 32:1527– Group A What has the model learned? (holisOc or modular level) Model AgnosOc
  • 15. Examples of Post-hoc Explana2ons 15 Henelius, A.; Puolama ̈ki, K.; and Ukkonen, A. 2017. Interpre?ng Classifiers through ADribute Interac?ons in Datasets. ICML Group A What has the model learned? (holis?c or modular level) Model Agnos?c
  • 16. General Concepts & Methods • Rule Sets • Sensi+vity Analysis • Induc+ve Logic/Programming • Recently: • Counterfactuals • Adversarial approaches • Game theory 16
  • 19. Open Challenges • No formal agreed upon defini1on • The no1on of interpretability seems to be an ill-defined term? • Having agreed upon defini1on avoids reinven1ng the wheel • Easier to built upon and contribute to prior work • Rigorous, agreed upon evalua1on metrics • Clear dis1nc1on of human vs. machine based evalua1on metrics • Provide a clear picture of what is working and what needs improvement 19
  • 20. Open Challenges • Stochas(c nature of the models, different random seeds lead to different outcomes for the same models • P Henderson, R Islam, P Bachman, J Pineau, Deep Reinforcement Learning That MaAers, AAAI 20018 • Models are built on assump(ons à = f( ) • When do they break and how? 20
  • 21. Open Challenges • Humans are great storytellers/story makers • Memory championship à Method of loci • Often humans create stories from small indications which rely upon in order to build explanations • These explanations might not have any relation with the underlying actual model • How to avoid specific cognitive biases? • Framing effect • Focusing effect • Illusory correlation 21
  • 24. Open Challenges • Saliency maps can be misleading (Olah et al., 2018) • Models are uncalibrated • Need for more transparent approaches • Bringing another to interpret the exisCng 24
  • 25. References 1. Bodenhofer, Ulrich and Bauer, Peter. Towards an Axiomatic Treatment of Interpretability, Fuzzy Systems, 2000. 2. Olah, Chris and Satyanarayan, Arvind. The Building Blocks of Interpretability, Distill.pub, 2018. 3. Zadrozny, Bianca and Elkan, Charles. Obtaining calibrated probability estimates from decision trees and naive bayesian classifiers. In ICML, pp. 609–616, 2001. 4. Zadrozny, Bianca and Elkan, Charles. Transforming classifier scores into accurate multiclass probability estimates. In KDD, pp. 694–699, 2002. 5. Naeini, Mahdi Pakdaman, Cooper, Gregory F, and Hauskrecht, Milos. Obtaining well calibrated probabilities using bayesian binning. In AAAI, pp. 2901, 2015. 6. Platt, John et al. Probabilistic outputs for support vector machines and comparisons to regularized likelihood methods. Advances in large margin classifiers, 10(3): 61–74, 1999. 7. Guo, Chuan and Pleiss, Geoff and Sun, Yu and Weinberger, Kilian Q. On Calibration of Modern Neural Networks. In ICML 2017. 25